A Survey on Federated Learning for Resource-Constrained IoT Devices

被引:357
作者
Imteaj, Ahmed [1 ,2 ]
Thakker, Urmish [3 ]
Wang, Shiqiang [4 ]
Li, Jian [5 ]
Amini, M. Hadi [1 ,2 ]
机构
[1] Florida Int Univ, Knight Fdn Sch Comp & Informat Sci, Miami, FL 33199 USA
[2] Florida Int Univ, Sustainabil Optimizat & Learning Interdependent N, Miami, FL 33199 USA
[3] SambaNova Syst, Deep Learning Res, Palo Alto, CA 94303 USA
[4] IBM TJ Watson Res Ctr, Yorktown Hts, NY 10598 USA
[5] SUNY Binghamton, Dept Elect & Comp Engn, Binghamton, NY 13902 USA
关键词
Edge computing; Computational modeling; Internet of Things; Data models; Servers; Collaborative work; Training; Convergence; federated learning (FL); global model; local model; on-device training; resource-constrained Internet-of-Things (IoT) devices; PRIVACY; OPTIMIZATION; INTERNET; COMMUNICATION;
D O I
10.1109/JIOT.2021.3095077
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Federated learning (FL) is a distributed machine learning strategy that generates a global model by learning from multiple decentralized edge clients. FL enables on-device training, keeping the client's local data private, and further, updating the global model based on the local model updates. While FL methods offer several advantages, including scalability and data privacy, they assume there are available computational resources at each edge-device/client. However, the Internet-of-Things (IoT)-enabled devices, e.g., robots, drone swarms, and low-cost computing devices (e.g., Raspberry Pi), may have limited processing ability, low bandwidth and power, or limited storage capacity. In this survey article, we propose to answer this question: how to train distributed machine learning models for resource-constrained IoT devices? To this end, we first explore the existing studies on FL, relative assumptions for distributed implementation using IoT devices, and explore their drawbacks. We then discuss the implementation challenges and issues when applying FL to an IoT environment. We highlight an overview of FL and provide a comprehensive survey of the problem statements and emerging challenges, particularly during applying FL within heterogeneous IoT environments. Finally, we point out the future research directions for scientists and researchers who are interested in working at the intersection of FL and resource-constrained IoT environments.
引用
收藏
页码:1 / 24
页数:24
相关论文
共 221 条
[1]   Mobile Edge Computing: A Survey [J].
Abbas, Nasir ;
Zhang, Yan ;
Taherkordi, Amir ;
Skeie, Tor .
IEEE INTERNET OF THINGS JOURNAL, 2018, 5 (01) :450-465
[2]  
Agarwal A., 2020, FEDERATED RESIDUAL L
[3]  
Agarwal N, 2018, ADV NEUR IN, V31
[4]  
Agrawal R, 2000, SIGMOD REC, V29, P439, DOI 10.1145/335191.335438
[5]  
Ammad-Ud-Din M., 2019, ARXIV190109888
[6]  
[Anonymous], 2018, arXiv
[7]  
[Anonymous], 2013, Cidr
[8]  
Assran Mahmoud, 2018, STOCHASTIC GRADIENT
[9]  
Banbury C., 2021, MICROETS NEURAL NETW
[10]  
Banbury C. R., 2021, Benchmarking TinyML systems: Challenges and direction